michaelmei/bigflow

A Python framework for BEAM data pipeline on GCP.

BigFlow

Documentation

1. What is BigFlow?
2. Getting started
3. Installing Bigflow
4. Help me
5. BigFlow tutorial
6. CLI
7. Configuration
8. Project structure and build
9. Deployment
10. Workflow & Job
11. Starter
12. Technologies
13. Development

Cookbook

• Monitoring
• Automated end-to-end testing
• Dockerized, GPU based ML prediction process

What is BigFlow?

BigFlow is a Python framework for data processing pipelines on GCP.

The main features are:

• Dockerized deployment environment
• Powerful CLI
• Automated build, deployment, versioning and configuration
• Unified project structure
• Support for GCP data processing technologies — Dataflow (Apache Beam) and BigQuery
• Project starter

Getting started

Start from installing BigFlow on your local machine. Next, go through the BigFlow tutorial.

Installing BigFlow

Prerequisites. Before you start, make sure you have the following software installed:

1. Python >= 3.8
2. Google Cloud SDK
3. Docker Engine

You can install the bigflow package globally, but we recommend installing it locally with venv, in your project's folder:

python -m venv .bigflow_env
source .bigflow_env/bin/activate

Install the bigflow PIP package:

pip install bigflow[bigquery,dataflow]

Test it:

bigflow -h

Read more about BigFlow CLI.

To interact with GCP you need to set a default project and log in:

gcloud config set project <your-gcp-project-id>
gcloud auth application-default login

Finally, check if your Docker is running:

docker info

Help me

You can ask questions on our gitter channel or stackoverflow.